It’s the Christian Democrats’ Fault: Declining Political Identification in Chile, 1957–2012

It’s the Christian Democrats’ Fault: DecliningPolitical Identification in Chile, 1957–2012

PATRICIO NAVIA New York University/Universidad Diego PortalesRODRIGO OSORIO Universidad de Santiago de Chile/Universidad

Diego Portales

The widely observed decline in party identification around the world doesnot necessarily respond to the same causes everywhere. The decline mightbe associated with institutional weakness, a crisis of democratic representa-tion, the rise of personal politics, generational replacement or post-material-ist values, and it might not be uniformly distributed across all parties. In fact,while identification with some parties might sharply fall, other parties mightexperience a smaller decline or no decline at all. Using the case of Chile, weshow that identification with the largest party has declined—and fluctuatedin general—far more than identification with the other parties. To better un-derstand decline in identification with parties, attention should be paid tovariance in party identification within countries, as a good deal of thedecline might be explained by falling identification with the largest party.

When democracy was restored in 1990, 71.5 per cent of Chileans iden-tified with a political party. In late 2013, the figure was 26 per cent. Thatdecrease—larger than in industrialized democracies—has received explana-tions that range from a democratic crisis of representation to the normaliza-tion of politics after an authoritarian experience. However, two-thirds of

This article was partially funded by Chile’s Fondo Nacional de Desarrollo Científico yTecnológico (FONDECYT #1140072) and the Centre for Social Conflict and CohesionStudies (CONICYT/FONDAP #15130009). We thank the two anonymous reviewers fortheir enlightening comments and suggestions. We also thank Peter Siavelis and ManuelAntonio Garretón for their comments and suggestions at the 2013 Latin AmericanStudies Association Conference.

Patricio Navia, New York University/Universidad Diego Portales, 726 Broadway,Room 666, New York NY, 10003, [email protected] Osorio, Universidad de Santiago de Chile/Universidad Diego Portales, Ejercito333, Santiago, Chile, [email protected]

Canadian Journal of Political Science / Revue canadienne de science politique48:4 (December / décembre 2015) 815–838 doi:10.1017/S0008423915001067© 2016 Canadian Political Science Association (l’Association canadienne de science politique)and/et la Société québécoise de science politique

mailto:[email protected]

mailto:[email protected]

the fall in party identification can be explained by the decline in identificationwith the largest party, the centre-left Christian Democratic Party (PDC).Identification with the PDC has gone from 37.2 per cent in 1990 to 6.2 percent in 2014. Identification with other parties has fallen by far less (34.3 to19.4%)—in line with a decline observed in advanced industrial democracies(Dalton and Wattenberg, 2000: 26). Using polling data from before the dem-ocratic breakdown of 1973 and from after 1990, we show that political partyidentification with parties other than the PDC has fluctuated far less over timeand fallen less markedly in recent years.

In what follows, we discuss the concept of party identification and thecauses for its widely reported decline. We then review existing evidence ondeclining party identification in Chile. Next, we present our data to showthat the variance in party ID between 1957 and 2014 is primarily explainedby fluctuations in identification with the PDC. As the party system in Chilehas been historically aligned on a left-centre-right continuum, we alsoanalyze the evolution of ideological identification on three-thirds, as thethree-way division is regularly called. Though there is an evident declinein identification with the centre, identification with the left and right has re-mained more stable. We finish by analyzing the evolution of the determi-nants of party identification and identification with the PDC in selectedpre-electoral polls for the 1960 to 2014 period. We show that the supportbasis for the PDC has aged and narrowed in terms of ideology.

Party Identification

Since first introduced, the concept of party identification seeks to “charac-terize the individual’s affective orientation to an important group-object”(Campbell et al., 1960: 121). The Michigan School considered party iden-tification a central determinant of voting preferences: “No single datum cantell us more about the attitude and behavior of the individual as presidentialelector than his location on a dimension of psychological identification ex-tending between the two great parties” (Campbell et al., 1960: 142–43).

Though initially used to predict voting patterns, party identificationsubsequently became a dependent variable. The focus was placed on thedeterminants of party identification more than on its consequences(Fiorina, 1981). Where Campbell and colleagues “spoke of a stable, affec-tive, and identity-based party identification of almost exogenous status,Fiorina saw a potentially volatile, rational, policy-based party identificationwith a clear endogenous position in the funnel of causality” (Holmberg,2007: 559).

Although it remains a controversial concept (Holmberg, 2007), partyidentification has been used as a proxy for institutionalization (Daltonand Weldon, 2007), party volatility (Drummond, 2006; Roberts and

816 PATRICIO NAVIA

Wibbels, 1999) and fragmentation (Coppedge, 1997). Support for parties isalso one of the indicators to assess party alignment and dealignment (Daltonet al., 1984).

The emergence of post-materialist values (Inglehart, 2007), the rise ofthe personal vote (Cain et al., 1987) and single-issue voters (Lau andRedlawsk, 2006) have diminished the importance of long-term determi-nants of voting preferences. The decline of social class as a determinantof political identification has also reduced the importance of party identifi-cation as an explanatory variable (Knutsen, 2007). Finally, the focus on theeconomic vote has conditioned the effect of party identification on votingbehaviour (Duch and Stevenson, 2008, Erikson et al., 2002; MacKuenet al., 1989).

Be it a dependent or independent variable, party identification has beenon the decline since first identified in the 1960s (Dalton, 2004; Holmberg,2007). That decline was well established in industrialized democracies inthe latter half of the twentieth century (Dalton and Wattenberg, 2000)and in emerging democracies (Holmberg, 2007), including Latin America(Hagopian, 1998), where it has been used as evidence of a crisis of demo-cratic representation (Mainwaring et al., 2007). The cause behind decliningparty identification in Latin America has been associated with partisan deal-ignment which “has proceeded faster in more countries than has realign-ment” (Hagopian, 1998: 126). Most of these studies look at the aggregatedecline for specific or groups of countries rather than at how identificationwith individual parties within countries has evolved. By using the case ofChile, we underline how the decline in identification at the national levelhides a wider variance in the evolution of identification with individualparties.

Abstract. For different reasons, identification with political parties has been on the decline ev-erywhere. Using polling data, we show that political identification in Chile increased in times ofpolarization before 1973 and in the transition to democracy in 1990 and fell under democratic con-solidation. The decline comes primarily from falling identification with the largest party, the centristChristian Democratic Party (PDC). Assessing the determinants of party identification, we reportthat while in the 1960s the PDC had a wide appeal as it attracted voters identified with the left,centre and right, in recent years the PDC attracts support from centrists and older people.

Résumé. Pour différentes raisons, l’identification partisane est en diminution partout. À partir dedonnées de sondages, nous démontrons que l’identification partisane au Chili a augmenté durant lapériode de polarisation politique antérieure à 1973, ainsi que durant la période de transitiondémocratique des années 1990, avant de diminuer durant la période de consolidationdémocratique. Le principal facteur qui explique ce déclin est la chute de l’identification avec leplus important parti politique, le Parti de la Démocratie chrétienne (PDC) situé au centre del’échiquier politique. L’analyse des déterminants de l’identification partisane révèle que durantles années 1960, le PDC attirait des électeurs de gauche, de centre et de droite alors que depuisles dernières années, il attire essentiellement des électeurs de centre et plus âgés.

Party Identification in Chile

Using electoral results, Carreras and colleagues find ample regional evidencefor a dealignment, but note that Chile and Honduras stand out as having“stable partisan alignment[s]” (2013: 13). The Chilean party system hasbeen compared to European party systems in its cleavage-based structure(Coppedge, 1998). Valenzuela argues that “forming a new party or runningas an independent… has been a somewhat risky strategy given the identifica-tion of a considerable portion of the electorate with the major party labels(1995: 4).”

Yet, using polling data, Bargsted and Somma find a “generalizedprocess of dealignment” in Chile (Bargsted and Somma, 2013: 11). Infact, Chile stands among the countries with the lowest levels of party iden-tification (Latinobarometro, 2013). The Latin American Public OpinionProject (LAPOP) 2010 report indicated that “except in Haiti, no othercountry in the region presents such a drastic decline in such a shortperiod of time” (Luna and Rosenblatt, 2012: 170). In the 2012 LAPOPreport, Chilean parties are noted “among those with the least sympathy”(Luna et al., 2012: 81) among all the countries included in the report.Reporting a decline in party identification across all age and incomegroups, Segovia notes that more Chileans identify with parties and on theideological scale on election years (2009).

Noting that identification with parties decreased from 70 per cent whendemocracy was restored to 40 per cent in 2009, Morales (2012) argues that“low volatility and low identification is exceptional for the theory of [party]institutionalization” and proceeds to define Chile, following Zucco (2010),as a “hydroponic party system” (Morales, 2012: 85), suggesting that partiesare strong but have shallow roots in society.

Discontent and malaise with the new democratic system was reported inthe 1990s (Moulián, 1997; PNUD, 1998; Riquelme, 1999). In 1998, Huneeusreported a “decline in party identification… which results in an increase inthe volatile vote” (1998: 4), associating it with the way in which the transitionto democracy occurred. While some have noted that “low- and medium-educated citizens and youth (with the exception of highly educated youth)are the groups that present lower levels of partisan attachment” (Luna andAltman, 2011: 19), others have claimed that declining levels of party identi-fication evidence a weak and insufficiently representative democracy (Atriaet al., 2013; Mayol, 2012; Portales, 2000; Winn 2004).

The decline in party identification has been linked with a decline inidentification on the left-centre-right scale (Huneeus, 2005: 72–73). Infact, “the declining meaning of ideological currents is matched by an in-creasing refusal to identify with any of the ideological currents [left,centre-left, centre, centre-right, right, and independent]” (Luna andAltman, 2011: 14–15).

818 PATRICIO NAVIA

These two conflicting views on party identification decline result fromthe fact that overall party identification has fallen primarily because of asteep decline on party identification with the PDC. In the early 1990s,when identification with the PDC increased, overall party identificationalso increased, but identification with other parties did not vary signifi-cantly. To be sure, identification with other parties has also fluctuated, butfar less than identification with the PDC. Detecting this trend in the mid-1990s, López and Martínez reported that PDC sympathizers do not easilyswitch allegiance to other parties (López and Martínez, 1999: 172–75).

The discrepancy on whether Chile has seen alignment or dealignmentresults from the fact that these studies treat party identification as a variablethat affects all parties equally. In showing that declining party identificationneed not uniformly affect all parties, we contribute to the discussion onwhat explains falling identification with parties and on the determinantsof party dealignment in emerging democracies.

Methodology

As the Chilean party system is usually analyzed on identification with indi-vidual parties and on a left-centre-right scale, we also use the traditionaldual focus on parties and “thirds.” We look longitudinally at the evolutionof party and ideological identification and use a cross-sectional analysis toexplore the determinants of that identification in three different moments.

We rely on polls conducted between 1957 and 1973 in Santiago bysociologist Eduardo Hamuy (Navia and Osorio, 2015). In 20 of thosepolls—conducted between October 1957 and February 1973—the ques-tionnaires inquired about identification with the left, centre, right or none.In 20 polls, conducted between 1961 and 1973, Hamuy inquired about iden-tification with political parties. Often, Hamuy also asked about identifica-tion with political parties and with the left, centre or right (not on theusual 1–10 scale). For the post-1989 period, we use polls conducted byCentro de Estudios Públicos (CEP) between 1990 and 2014. These pollsused probabilistic national samples. Normally, CEP conducts two pollsper year. Altogether, there are 58 post-1989 national polls where respon-dents were asked to self-identify with the left, centre or right and with po-litical parties. The pre-1973 polls were conducted only in Santiago. Thepost-1989 polls were national samples. Though they are not fully compara-ble, the pre-1973 Hamuy polls are the only survey data available to studypolitical party and ideological identification.

For the cross-sectional analysis, we conduct multinominal logit regres-sions using the Hamuy and CEP polls for years immediately preceding anelection year. Unlike Torcal and Mainwaring (2003) who use polling datafrom 1973 and 1995 to assess changes in the social bases of party

It’s the Christian Democrats’ Fault 819

competition, we avoid using years of unusually high levels of polarization.As we discuss below, 1973 was the year with the highest level of party iden-tification, while 1995 saw a steep decline in party identification. Since partyidentification increases in election years, we select years that are compara-ble in terms of the electoral cycle. Thus we analyze the determinants ofparty identification and identification with the PDC in the pre- (presidential)election years of 1969, 1992, 2004 and 2012. We exclude the 1989 and1999 presidential elections because they were held under authoritarianrule (1989) or non-concurrently with legislative elections (1999).

The Evolution of Political Identification in Chile

In the twentieth century, the Chilean party system constituted along a left-centre-right alignment (Collier and Sater, 1996; Valenzuela, 1978;Valenzuela and Valenzuela, 1976). Parties were described as having deeproots in society, with a well-defined class-based structure (Drake, 1978;Scully, 1992). The incorporation of the urban working class explains therise of left-wing Communist (PC) and Socialist (PS) parties (Angell,1972; Drake, 1978). The incorporation of the rural poor in the 1950s rede-fined the political landscape, favouring the PDC and weakening the PR(Fleet, 1985; Grayson, 1968; Scully, 1992; Valenzuela, 1995). The PDC-led agrarian reform increased political polarization in the countryside andweakened the landowning electoral base of right-wing parties (Loveman,1976; Scully, 1992; Valenzuela, 1978).

As the new leading centrist party, the PDC failed to facilitate the for-mation of multi-party government coalitions and thus a polarized multi-party system emerged in the late 1960s (Valenzuela, 1978, 1990). Basedon studies of party elites, polarization of the party system at the elitelevel is said to have contributed to the democratic breakdown(Valenzuela, 1978), but others have blamed the extreme parties for polariz-ing the system (Moulian, 1993; Scully 1992). Using data from three polls,the claim of increasing polarization has been challenged (Prothro andChaparro, 1974). Declining identification with the PDC coincided withthe breakdown of democracy in the early 1970s and with democratic con-solidation since the early 1990s.

The two traditional right-wing parties, Liberal and Conservative, badlylost the 1965 legislative election and merged into the National Party (PN) in1967 (Correa Sutil, 2005; Valenzuela, 1995). Thus, the pre-1973 partysystem comprised the Liberal and Conservative parties (and later the PN)on the right, the PDC and PR on the centre and the PS, and PC on theleft (Moulian. 1993).

With the restoration of democracy in 1990, several pre-1973 partiesmade a comeback (Angell 2003; Siavelis, 2002). Drawing on continuities

820 PATRICIO NAVIA

in party organization and electoral preferences before 1973 and after 1990,the persistence of the three-thirds alignment has been noted (Garretón,1989; Valenzuela, 1995; Valenzuela, 1999; Valenzuela and Scully, 1997;Valenzuela et al., 2007). Others have identified the emergence of a new au-thoritarian/democratic divide in the political party system (Bonilla et al.,2011; Montes et al., 2000; Tironi and Agüero, 1999; Torcal andMainwaring, 2003). The electoral system has also been identified as abarrier that prevents the three-thirds divide from materializing (Navia,2005; Siavelis, 2004). The across-the-board two-seat proportional represen-tation district arrangement, known as the binominal system, induced the for-mation of two coalitions that challenged the old three-thirds alignment(Magar et al., 1998).

The PDC became the most important party in the mid-1960s and hasremained so until today—in terms of party identification, though it is notthe party with the highest vote share. Yet, the electoral bases of PDCchanged during the dictatorship (Torcal and Mainwaring, 2003;Valenzuela and Scully, 1997). In post-Pinochet democracy, Bonilla andcolleagues (2011) and Dow (1998) find PDC sympathizers self-identifyas left-of-centre and report that identification with the PDC and left-wingparties is strongly associated with opposition to the Pinochet regime.Identification with the centre on a left-right scale and religion, in urbanareas, is an important predictor of support for the PDC (Morales andPoveda, 2007). Huneeus (2003) showed declining identification with thePDC among the youth from 1990 to 2001.

Overall party identification has evolved substantially over time.Figure 1 shows the evolution in party id from 1957 until 2014 (excludingthe Pinochet period), with and without the PDC. When we exclude identi-fication with the PDC, evolution in identification with all other parties fluc-tuates less.

To analyze the evolution in identification with parties other than thePDC, we separate the analysis in two figures. Figure 2 includes the threeleft-wing parties, the PS, PC and Party for Democracy (PPD), and two cen-trist parties, the PDC and PR. Four parties—PDC, PR, PS, PC—have beenactive uninterruptedly since the mid-1950s. With the exception of the 1970–1973 period, identification with the PS, PC and PR has remained below 10per cent. The PS enjoyed a peak over 20 per cent under Allende. After 1990,though lower than its pre-1973 peak, identification with the PC and PS wasat similar levels as in the late 1950s and early 1960s. Identification with thePR shows a decline in the late 1950s and stability since the early 1970s.Identification with the PPD has fluctuated between 5 and 10 per cent.

Identification with the PDC fluctuated significantly, between 10 and50 per cent. It reached its peak when Eduardo Frei became president in1964. After 1990, support for the PDC remained high under the presiden-cies of PDC Patricio Aylwin (1990–1994) and Eduardo Frei Ruiz-Tagle


(1994–2000). Decline in identification with the PDC began in the mid-1990s and continued after 2000, when the PDC was a minority partner inthe Concertación, under presidents Ricardo Lagos (PPD, 2000–2006) andMichelle Bachelet (PS, 2006–2010).

Figure 3 shows identification with right-wing parties from 1961 to2014. For before 1966, we combined the identification with theConservative and Liberal parties. For after 1990, we report identificationwith Unión Demócrata Independiente (UDI), Renovación Nacional (RN)and the combined identification with the two right-wing parties.

Identification with the right shows an upward trend since the 1960s.Identification with the PN grew to 18 per cent in the first year of Allendebefore declining by 1973. After 1989, identification with right-wingparties has reached peaks of more than 20 per cent after the 1999 and2005 presidential elections. The peak with UDI happened after 1999 andwith RN in 2005, years when each party had popular presidential candi-dates. With the exception of the period around 1999, more people identifywith RN than with UDI. The combined identification with the two rightparties shows fluctuation—not just decline—over time.

The highly polarized party system identified as one of the causes of thebreakdown of democracy is not reflected in the patterns of political partyidentification shown in Figures 2 and 3. There was an increase in

FIGURE 1.Identification with political parties, with and without PDC 1957–2014

Source: Authors, with data from Eduardo Hamuy polls, 1957–1973 and CEP polls,1990–2014

822 PATRICIO NAVIA

identification with the PS and PC before 1973, but at the height of politicalpolarization, fewer people identified with left-wing and right-wing partiesthan with the PDC. Considering that it was a period of high political polar-ization, there was high fluctuation in the level of identification with thePDC, and to a lesser extent with right-wing parties before 1973, whilesupport for left-wing parties was more stable (with the exception of theAllende period).

The claim that political parties have always enjoyed high levels ofidentification is not supported by public opinion data. The overall trendpoints to higher volatility in identification with the PDC and more stableidentification with the other parties. Though some parties have experiencedupticks in times of polarization, as in the early 1970s and during the tran-sition to democracy in the early 1990, all other parties show far less variancein their identification during the period than the PDC.

Identification on the Left-Centre-Right Scale

The evolution in identification on the left-centre-right scale can shed lighton why identification with the PDC evolved differently than identificationwith other parties. In Chile, polls normally ask individuals about their

FIGURE 2.Identification with Centre and Left Political Parties, 1961–2014



political identification with the left, centre or right rather than on a 1–10scale as is normally done in other countries. Both the Hamuy and CEPpolls ask specifically if Chileans identify with the left, centre or right. Inthat regard, pre-1973 and post-1990 polls are comparable in assessing po-litical identification.

In the early 1960s, close to 30 per cent identified with the left andcentre, and around 25 per cent did so with the right, as shown inFigure 4. Identification with the left reached almost 50 per cent in 1972.Identification with the centre remained at around 30 per cent, while identi-fication with the right was steady near 20 per cent. As Chilean politicsbecame more polarized, identification with the right did not grow, but iden-tification with the left increased drastically.

By 1972, the three-thirds had evolved into a two-way polarizing dividebetween a leftist half and a centre-right half in Congress. Political polariza-tion is associated with the weakening of the centre (Sartori 1976). Yet,before 1973, there was a growth in identification with the left, a smaller in-crease in identification with the right, a large decline in identification withthe PDC, but identification with the centre remained stable.

In the early 1990s, identification with the centre was at similar levels asbefore 1973. After falling below 10 per cent in 1997, it has since fluctuatedbetween 10 and 15 per cent. Identification with the left and right has

FIGURE 3.Political Identification with rightwing parties, 1961–2014


824 PATRICIO NAVIA

fluctuated more, with the highest support for the left a few years after thetransition. A downward trend since then has brought support for the leftto around 20 per cent. Identification with the right was lowest in theearly 90s, with less than 15 per cent. Figure 4 shows a steady decline inidentification with the three thirds since 1990. After being the most stablebefore 1973, the centre was the first third to see a decline after the transition.

Figure 5 plots the identification with the centre and vote share for thePDC between 1957 and 2014. Before 1973, support for the centre waslower than the PDC vote share. While in 1973 less than 30 per cent sympa-thized with the centre, the vote share for centrist parties was close to 40 percent. Higher vote share than poll support for centrist parties before 1973might result from the rapid growth of the PDC in rural areas.Unfortunately, the Hamuy polls in our study were only conducted inSantiago. Since 1989, the PDC vote share has always been higher thanthe level of identification with the centre. Thus, the PDC electoral baseextends beyond those who identify only as centrists.

Studies on electoral behaviour and political preferences in Chile havenormally assumed that voters who identify with political parties also iden-tify with the corresponding third on the left-centre-right scale (Morales andPoveda, 2007: 149; Briceño and Navia, 2008). However, as Table 1 shows,

FIGURE 4.Political Identification with the Left-Centre-Right in Chile, 1957–2014



those who identify with the PDC in three polls conducted a year before the1970, 1993 and 2013 presidential elections shows a far less than perfectoverlap between identification with political parties and on the left-centre-right scale. Though overall identification with the PDC variedfrom 43.2 per cent in 1969 to 40.4 per cent in 1992 and 5.3 per cent in2012, those who identified with the party were consistent in the way theyaligned themselves on the ideological scale. Close to 20 per cent of PDCsympathizers identified with the right, whereas around 50 per cent indicated

FIGURE 5.Political Identification with Centre and votes for the PDC, 1957–2014

Source: Authors, with data from Eduardo Hamuy polls, 1957–1973, CEP polls,1990–2014 and Cruz-Coke (1984)

TABLE 1.Identification with PDC by identification with thirds (per cent)

1969 1992 2012

Right 18.7 19.9 17.7Centre 18.7 28.3 29.9Left 50.4 47.3 54.5None 12.3 4.5 3.9Total 100 100 100per cent identified with PDC 34.2 40.4 5.2

Source: Hamuy Poll (1969) and CEP Polls (1992 and 2012)

826 PATRICIO NAVIA

an identification with the left. Those identified with the centre increasedfrom less than 20 per cent in 1969 to close to 30 per cent in 1992 and2012. Thus, political party ID and ideological self-positioning should notbe treated the same way, at least for those identified with the PDC. Whileidentification with the PDC declined over time, the breakdown of PDCsympathizers continued to reflect a wide ideological identification on theleft-centre-right scale.

Since 1990, the dominant role played by the centre-left Concertacióncoalition and the centre-right Alianza has made it more difficult to usethe three-thirds to describe the Chilean electorate. As Torcal andMainwaring (2003) and Bonilla and colleagues (2011) have argued,parties form coalitions that, in turn, align along an authoritarian-democraticdivide. However, identification with coalitions has also fallen in line withdeclining party identification. In 1994, when CEP polls first asked for iden-tification with coalitions, 47 per cent identified with the Concertación and19 per cent with the Alianza. By 2012, those numbers had fallen to 22 percent and 14 per cent respectively, showing a steeper decline in the coalitionthat includes the PDC.

The decline in identification with the PDC and the centre is correlatedwith an increase among those who do not identify with parties or on thescale. Figure 6 shows those who identify with “none,” both on the left-centre-right scale and with political parties. There is a U-shape trend.There were low levels of party and ideological identification in the late1950s and early 1960s, but those choosing “none” decreased starting inthe late 1960s. After 1990, it began to increase rapidly. The “none” ishigher on the party than on the left-centre-right scale. When identificationwith the thirds declines, so does identification with parties. We do notknow which dimension of identification drives the other. It might be thatthey are both indicators of a broader notion of political identification.

In the October 1957 poll, 58.5 per cent identified with no politicalsector. The fact that President Carlos Ibáñez (1952–1958) was an indepen-dent might explain the low identification with the thirds, but even if weexclude those who identified with “other,” 29.6 per cent chose not to iden-tify on the left-centre-right scale. Presidential elections are associated withan increase in polarization. Consistently, in the first poll conducted afterFrei’s 1964 victory, those not identified on the ideological scale fell to10 per cent, but by 1965, the number increased again to 25 per cent. Afterthe 1970 election, levels of identification remained high. The low identifi-cation with “none” in February of 1973 (8.6%), the last poll conductedbefore the democratic breakdown, reflected high polarization. After democ-racy was restored, identification with “none” began to experience a steadyincrease. Though there was polarization at times of turmoil in the late 1960sand transition to democracy in the early 1990s, in politically normal times,about a fourth of Chileans have consistently failed to identify politically.


Determinants of Identification with the PDC

Since the evolution in support for other parties has varied far less, we focusin this last section on changes in identification with the PDC. To assess theevolution in support for the PDC, we look at the determinants for identifi-cation with the PDC before and after 1973. Since the polls were different intheir design and scope, we separate the two periods. We estimated threemultinominal regression models. The dependent variable is identificationwith political parties. We grouped identification in four categories, identifi-cation with the PDC, identification with leftist parties (those parties to theleft of the PDC), identification with right-wing parties and none. We useidentification with the PDC as our reference category.

Following classic studies on party identification, we include as controlvariables long, medium and short-term determinants of party identification.We use level of education, socio-economic status, political identification onthe thirds (left-centre-right scale), approval for the president, and socio-tropic perceptions. Sociotropic perceptions on the economy, followingErikson and colleagues (2002), are coded from ‘−1’ (worse) to ‘1’(better). As control variables, we include gender and age. Table 2 reportsthe findings for the Hamuy poll in 1969 and Table 3 reports the findingsfor the CEP polls in 2002 and 2012.

FIGURE 6.Political Identification with PDC and “None”, 1957–2014


828 PATRICIO NAVIA

Table 2 confirms that the ideological bases of the PDC were not restrict-ed to those who identified with the centre. Those not identified on the ideo-logical third (right, centre or left) were more likely to identify with no party,but they were also more likely to identify with the PDC than with left-wingparties. This is consistent with findings that report PDC sympathizers to

TABLE 2.Multinomial logit regression on PDC identification, 1969

Left-wing parties Right-wing parties None

Female 0.107 −0.184 0.135(0.333) (0.462) (0.319)

Age 0.00327 0.257a −0.0426(0.0683) (0.0963) (0.0661)

EducationPrimary*

Secondary 0.196 0.495 0.312(0.366) (0.517) (0.350)

Tertiary 0.0697 1.049 1.152c

(0.785) (0.948) (0.657)Socio-economic statusLow*

Medium −0.0167 1.047c 0.268(0.370) (0.604) (0.363)

High 0.628 2.632a 0.801(0.554) (0.758) (0.511)

Identification with thirdsCentre*

None −0.958c −0.556 1.015b

(0.496) (0.966) (0.434)Left −1.387a 0.00361 −1.233a

(0.397) (0.651) (0.406)Right −1.619a 1.955a −1.101b

(0.542) (0.700) (0.530)Presidential approval −1.818a −1.739a −0.685b

(0.366) (0.494) (0.345)Sociotropic evaluation −1.263a −1.449a −1.185a

(0.300) (0.408) (0.283)Constant 0.967c −3.639a −0.119

(0.514) (0.985) (0.535)

Observations 379 379 379

The reference category for the dependent variable is identification with the PDC*Reference categoryStandard errors in parenthesesap < 0.01bp < 0.05cp < 0.1Source: Hamuy Poll (April, 1969)


TABLE 3.Multinomial logit regression on PDC identification, 1992 and 2012

1992 2012

Left-wing parties Right-wing parties None Left-wing parties Right-wing parties None

Female −0.238c −0.199 −0.309 0.340 −0.471 0.0289(0.136) (0.153) (0.194) (0.278) (0.347) (0.260)

Age −0.00551 −0.00207 0.00300 −0.0340a −0.0275a −0.0294a

(0.00470) (0.00519) (0.00670) (0.00866) (0.0105) (0.00810)EducationPrimary*

Secondary 0.0435 0.252 0.49a −0.0486 0.263 −0.158(0.164) (0.180) (0.231) (0.332) (0.416) (0.307)

Tertiary 0.436b 0.556b 0.904a 0.732 0.886 0.886(0.210) (0.237) (0.308) (0.584) (0.677) (0.557)

Socio-economic statusLow*

Medium −0.0667 −0.0172 0.334 0.170 0.312 0.240(0.149) (0.170) (0.216) (0.286) (0.368) (0.268)

High −0.0845 0.654c 0.209 −0.520 −1.037 −0.171(0.344) (0.340) (0.504) (1.492) (1.370) (1.168)

Identification with thirdsCentre*

None 0.475 −0.559c 3.495a 0.992c 0.450 4.077a

(0.342) (0.336) (0.254) (0.584) (0.812) (0.507)Left 1.529a −2.476a −1.012a 0.984a −2.228a −0.161

(0.182) (0.309) (0.309) (0.333) (0.811) (0.307)Right 0.343 0.834a −0.0791 −0.183 2.432a 0.392

(0.235) (0.164) (0.282) (0.489) (0.471) (0.399)

830PATRICIO

NAVIA

Presidential approval −0.552a −1.219a −1.161a −0.0427 0.711c 0.0795(0.150) (0.158) (0.200) (0.356) (0.406) (0.322)

Sociotropic evaluation −0.130 −0.165 −0.121 0.0827 0.499c 0.287(0.109) (0.123) (0.158) (0.218) (0.267) (0.205)

Constant −0.827a 0.103 −1.645a 1.588b 0.300 2.379a

(0.304) (0.308) (0.418) (0.644) (0.778) (0.596)

Observations 1,779 1,779 1,779 1,340 1,340 1,340

The reference category for the dependent variable is identification with the PDC*Reference categoryStandard errors in parenthesesap < 0.01bp < 0.05cp < 0.Source: CEP Polls (November, 1992 and November, 2012)

It’sthe

Christian

Dem

ocrats’Fault

831

self-locate left of centre in the ideological scale after 1990 (Bonilla et al.,2011; Dow, 1998). Those identified with the left were more likely to iden-tify with the PDC than with left-wing parties. That reflected both the hetero-geneity of left-wing parties and the reformist rhetoric of the PDC (after all,the PDC came to power in 1964 promising a Revolution in Liberty). Thus,those who self-defined as leftist were inclined to identify with the PDC justas those who identified with the PDC were more likely to define themselvesas leftists.

Those who self-defined as centrists were strongly identified with right-wing parties. They were less likely to identify with the PDC than with right-wing parties but they were also more likely to identify with the PDC thanwith no party. Thus, those self-defined as rightists who did not adhere toa right-wing party were also potential PDC sympathizers in 1969. That re-flects the broad base appeal of the PDC that attracted support from amongthose self-defined as leftist and rightist.

Not surprisingly, Table 2 also shows that in 1969, those who identifiedwith parties other than the PDC or with no party at all were less likely toapprove of President Frei and less likely to have a positive economicoutlook. In terms of socio-economic strata, the electoral base of the PDConly differed from those of right-wing parties. The base of PDC sympathiz-ers was not statistically different than the base of sympathizers of left-wingparties or those not identified with any party.

Table 3 shows two multinominal logit regressions for 1992 and 2012.Since political identification normally increases in elections, we chose theyear before a presidential election. The models reflect the change in thesupport base for the PDC over the period. As identification with the PDCdecreased, the basis of support for that party also changed.

In 1992, there were no age differences in the bases of support for thePDC with regards to left-wing and right-wing parties (and with those notidentified). In 2012, PDC sympathizers were systematically older thanthe rest. The evolution in identification with the PDC between 1969 and2012 shows the aging in identification with the PDC overtime. In the1960s, the PDC had stronger support among younger Chileans while by2012, its strongest support base was among older Chileans.

In terms of gender, more women identified with the PDC than left-wing parties in 1992, but that was no longer the case in 2012. The baseof support for the PDC was strongest among those with lower levels of ed-ucation. Those with higher educational levels were more likely to identifywith left-wing and right-wing parties than with the PDC. Those with highereducational levels were also more likely to identify with no party than withthe PDC. The PDC was also less likely than right-wing parties to attractsupport from people in higher socio-economic status.

In 1992, those not identified with the left, centre or right were stronglymore likely not to identify with any political party, but they were also more

832 PATRICIO NAVIA

likely to identify with the PDC than with right-wing parties. Those whoself-defined as leftist were obviously more likely to identify with left-wing parties but also more likely to identify with the PDC than withright-wing parties or with no party. In 2012, those not identified with theleft, centre or right were more likely to identify with no parties, but alsomore likely to identify with left-wing parties than with the PDC or right-wing parties. In 2012, those identified with the left continued to identifymore with the PDC than with right-wing parties. Those self-defined asrightist continued to identify primarily with right-wing parties and notwith the PDC in 2012.

In 1992, when PDC President Patricio Aylwin had high approval, pres-idential approval had a positive effect on identification with the PDC withrespect to other categories. In 2012, when the president was from the right-wing Alianza, presidential approval explained identification with right-wing parties when compared to the PDC. Unlike 1969, sociotropic percep-tions were not significant in 1992 or 2012.

After 1990, identification with the PDC was lower among rightiststhan among those identified with the centre. Something similar happenedwith those identified with the left. They were no longer more likely to iden-tify with the PDC. Those identified with none were also far less likely toidentify with the PDC than those identified with the centre. Unlike 1969,when the PDC drew support from all sectors, after 1990, the supportbasis for the PDC was restricted to those identified with the centre. Asthe percentage of those identified with the centre declined, support forthe PDC also declined.

Since the 1990s, PDC sympathizers have been more likely to identifywith the left than with the centre. However, in 2012, the support base for thePDC no longer showed a higher likelihood among leftists than centrists toidentify with the PDC. Over time, the PDC lost its support base both in theright and left and was only able to maintain its support basis among thoseself-defined as centrists.

Before 1973, the PDC attracted support from those identified with theright and left, but after 1990 the PDC obtained its support primarily fromthose identified with the centre. As the percentage of centrists declined,identification with the PDC also fell rapidly. Since those not identified onthe ideological scale have markedly increased since 1990 and the PDC—or any other party—has failed to attract support from that group, the elec-toral base for the PDC has continued to shrink. As compared to before1973, when the PDC also attracted support from the left and right, after1990, the inability of the PDC to attract support from those identifiedwith the left, right or none explains the rapid decline in identificationwith the PDC. Since the pre-1973 polls were only conducted in Santiago,we also estimated the models for the post 1990 polls using only Santiago


residents. The results did not differ from those using the entire nationalsample.

Conclusion

As in most other countries, identification with political parties in Chile hasdeclined in recent years. However, the evolution of identification withparties for the entire 1957–2014 period is more nuanced. In periods of po-larization, identification with parties increased while in periods of politicalnormalcy, it declined. Most of the decline can be explained by falling iden-tification with the PDC, the largest party. Identification with other partieshas fallen far less.

Identification on the left-centre-right scale should not be automaticallyequated with identification with political parties that ascribe to each ideo-logical third. In fact, the PDC received support from people identifiedwith the left, centre and right. Though the party self-defined as centrist(ideologically different from the socialist left) in the 1960s and as centre-left after the restoration of democracy, PDC sympathizers before 1973saw themselves primarily as leftists more than centrists or rightists. After1990, PDC sympathizers have primarily defined themselves as centrists.

Though party identification there has fallen as elsewhere in the world,the case of Chile shows that most of the decline is explained by falling iden-tification with one large party, the PDC. If the PDC is excluded, politicalidentification in Chile has declined far less, with periods of up and downassociated with elections and political polarization.

To be sure, political parties have remained, as Garretón noted, thebackbone of the political process in Chile (1989: xvi). Before 1973 andafter 1990, they have been central to the political evolution in thecountry. The data we present do not challenge abundant references to thehigh level of party institutionalization in Chile but simply introducenuances as we show that a large segment of society never identified withpolitical parties in polls. The fact that political parties have historically dom-inated electoral politics—with Chileans showing scant interest in support-ing independent candidates—underlines the importance of our findings.With the exception of the PDC, identification with parties was stable andlow but people continued to vote for candidates from traditional parties.Moreover, since identification on polls with the left-centre-right scale washistorically higher than identification with political parties, we concludethat ideology was an important determinant of political identification inChile.

The case of Chile can enlighten our understanding of declining partyidentification elsewhere. A drastic fall in the identification with a largeparty can lead to wrong generalizations about declining identification

834 PATRICIO NAVIA

with all parties. Even if identification with other parties remains constant,when identification with the largest party falls drastically, observersmight wrongly conclude that the country is experiencing an overall partyidentification dealignment. By analyzing the evolution in party identifica-tion by parties—and not just at the national level—richer conclusions canbe made about what explains the overall decline in party identification.

References

Angell, Alan. 1972. Partidos políticos y movimiento obrero en Chile. Mexico DF: EdicionesERA.

Angell, Alan. 2003. “Party Change in Chile in Comparative Perspective.” Revista de CienciaPolitica 23 (2): 88–108.

Atria, Fernando, Guillermo Larraín, José Miguel Benavente, Javier Couso, and AlfredoJoignant. 2013. El otro modelo. Del orden neoliberal al régimen de lo público.Santiago: Debate.

Bargsted, Matías A. and Nicolás M Somma. 2013. “Social cleavages and political dealign-ment in contemporary Chile, 1995–2009.” Party Politics. Doi: 1354068813514865.

Bonilla, Claudio, Ryan E. Carlin, Gregory J. Love and Ernesto Silva Méndez. 2011. “Socialor political cleavages? A spatial analysis of the party system in post-authoritarian Chile.”Public Choice 146 (1): 9–21.

Briceño, Renato and Patricio Navia. 2008. “El PDC y la votación de Ricardo Lagos en laspresidenciales de 1999.” Revista de Ciencia Política 28 (2):107–31.

Cain, Bruce E., John Ferejohn and Morris P. Fiorina. 1987. The Personal Vote: ConstituencyService and Electoral Independence. Cambridge: Harvard University Press.

Campbell, Angus, Philip E. Converse, Warren E Miller and Donald E Stokes. 1960. TheAmerican Voter. New York: Wiley.

Carreras, Miguel, Scott Morgenstern and Yen-Pin Su. 2013. “Refining the theory of partisanalignments Evidence from Latin America.” Party Politics. Doi: 1354068813491538.

Collier, Simon and William F. Sater. 1996. A History of Chile. 1808–1994. Cambridge:Cambridge University Press.

Coppedge, Michael. 1997. “District Magnitude, Economic Performance, and Party SystemFragmentation in Five Latin American Countries.” Comparative Political Studies inComparative International Development 30 (1):156–85.

Coppedge, Michael. 1998. “The Dynamic Diversity of Latin American Party Systems.” PartyPolitics 4 (4): 547–68.

Correa Sutil, Sofía. 2005. Con las riendas del poder. La derecha chilena en el siglo XX.Santiago: Editorial Sudamericana.

Dalton, Russell J. 2004. Democratic Challenges, Democratic Choices. The Erosion ofPolitical Support in Advanced Industrial Democracies. New York: Oxford UniversityPress.

Dalton, Russell J. and S. Weldon. 2007. “Partisanship and party system institutionalization.”Party Politics 13 (2):179–96.

Dalton, Russell and Martin Wattenberg. 2000. Parties without Partisans: Political Change inAdvanced Industrial Democracies. Oxford: Oxford University Press.

Dalton, Russell J., S.C. Flanagan, P.A. Beck and J.E. Alt. 1984. Electoral change in ad-vanced industrial democracies: realignment or dealignment? Princeton NJ: PrincetonUniversity Press.

Dow, J. K. 1998. “A Spatial Analysis of Candidates in Dual Member Districts: The 1989Chilean Senatorial Elections” Public Choice 97: 119–142.


Drake, Paul W. 1978. Socialism and Populism in Chile. 1932–1952. Urbana: University ofIllinois Press.

Drummond, Andrew J. 2006. “Electoral Volatility and Party Decline in WesternDemocracies: 1970–1995.” Political Studies 54 (3): 628–47.

Duch, Raymond M. and Randolph T. Stevenson. 2008. The Economic Vote. How Politicaland Economic Institutions Condition Election Results. New York: CambridgeUniversity Press.

Erikson, Robert S., Michael B. MacKuen and James Stimson. 2002. The Macro Polity.New York: Cambridge.

Fiorina, Morris. P. 1981. Retrospective Voting in American National Elections. New Havenand London: Yale University Press.

Fleet, Michael. 1985. The Rise and Fall of Chilean Christian Democracy. Princeton N.J:Princeton University Press.

Garretón, Manuel Antonio. 1989. The Chilean Political Process. Boston, Unwin Hyman.Garretón, Manuel Antonio. 2000. “Chile’s Election: Change and Continuity.” Journal of

Democracy 11 (2): 78–84.Grayson, George. 1968. El Partido Demócrata Cristiano Chileno. Santiago: Editorial

Francisco de Aguirre.Hagopian, Frances. 1998. “Democracy and political representation in Latin America in the

1990s: Pause, reorganization, or decline?” In Fault lines of democracy in post-transitionLatin America, ed. Felipe Agüero and Jeffrey Stark. Coral Gables Fl: North-South CenterPress, University of Miami.

Holmberg, Sören. 2007. “Partisanship Reconsidered.” In The Oxford Handbook of PoliticalBehavior, ed. R.J. Dalton and Hans-Dieter Klingemann. New York: Oxford UniversityPress.

Huneeus, Carlos. 1998. “Malestar y desencanto en Chile: Legados del autoritarismo y costosde la transición.” Papeles de Trabajo. Corporación Tiempo 2000 (63).

Huneeus, Carlos. 2003. “Un partido con un alto grado de institucionalización: el PDC enChile.” In Christian Democracy in Latin America, ed. Scott Mainwaring and TimothyR. Scully. Stanford: Stanford University Press.

Huneeus, Carlos. 2005. “Las coaliciones de partidos: ¿Un nuevo escenario para el sistemapartidista chileno?” Política 45: 68–86.

Inglehart, Ronald. 2007. “Post Materialist Values and the Shift From Survival to Self-Expression Values.” In The Oxford Handbook of Political Behavior, ed. Russell J.Dalton and Hans-Dieter Klingemann. New York: Oxford University Press.

Knutsen, Oddbjorn. 2007. “The Decline of Social Class?” In The Oxford Handbook ofPolitical Behavior, ed. Russell J Dalton and Hans-Dieter Klingemann. New York:Oxford University Press.

Latinobarometro. 2013. Informe Latinobarómetro 2013. Santiago: Latinobarómetro.Lau, Richard R. and David P. Redlawsk. 2006. How Voters Decide. Information Processing

During Elections Campaigns. New York: Cambridge University Press.López, Miguel Ángel and Gustavo Martínez. 1999. “Opinión pública y democracia: las

encuestas en la redemocratización chilena.” In La Caja de Pandora. El retorno de latransición chilena, ed. Amparo Menéndez-Carrión and Alfredo Joignant. Santiago:Planeta.

Loveman, Brian. 1976. Struggle in the Countryside: Politics and Rural Labor in Chile, 1919–1973. Blomington IN: Indiana University Press.

Luna, Juan Pablo, and David Altman. 2011. “Uprooted but Stable: Chilean Parties and theConcept of Party System Institutionalization.” Latin American Politics and Society 53(2): 1–28.

Luna, Juan Pablo and Fernando Rosenblatt. 2012. “¿Notas para una autopsia? Los partidospolíticos en el Chile actual.” In Democracia con partidos. Informe para la reforma a

836 PATRICIO NAVIA

los partidos políticos en Chile, ed. Francisco Javier Diaz and Lucas Sierra. Santiago:CEP/ CIEPLAN.

Luna, Juan Pablo, Sergio Toro, Nathalie Jaramillo-Brown, Valentina Salas and MitchellSeligson. 2012. Cultura política de la democracia en Chile y en las Américas, 2012:Hacia la igualdad de oportunidades. Santiago: LAPOP/Ponitificia UniversidadCatólica.

MacKuen, Michael B. Robert S. Erikson, and James A. Stimson. 1989. “Macropartisanship.”The American Political Science Review 83 (4): 1125–42.

Magar, Eric, Marc R. Rosenblum, and David Samuels. 1998. “On the absence of centripetalincentives in double-member districts. The case of Chile.” Comparative Political Studies31 (6): 714–39.

Mainwaring, Scott, Ana María Bejarano and Eduardo Pizarro. 2007. The Crisis ofDemocratic Representation in the Andes. Stanford: Stanford University Press.

Mayol, Alberto. 2012. No al lucro. De la crisis del modelo a la nueva era política. Santiago:Random House Mondadori.

Montes, Juan E., Scott Mainwaring and Eugenio Ortega. 2000. “Rethinking the Chilean PartySystems.” Journal of Latin American Studies 32 (3) :795–824.

Morales, Mauricio. 2012. “The Concertación’s Defeat in Chile’s 2009–2010 PresidentialElections.” Latin American Politics and Society 54 (2): 79–107.

Morales, Mauricio and Antonio Poveda. 2007. “El PDC: bases electorales, determinantes deadhesión e impacto en las votaciones de R. Lagos y M. Bachelet.” Estudios Públicos107: 129–65.

Moulian, Tomás. 1993. La forja de ilusiones. El sistema de partidos 1932–1973. Santiago:FLACSO-ARCIS.

Moulián, Tomás. 1997. El Chile Actual. Anatomía de un mito. Santiago: LOM-Arcis.Navia, Patricio. 2005. “Transformando votos en escaños: leyes electorales en Chile, 1833–

2003.” Política y Gobierno 12 (2): 233–76.Navia, Patricio and Rodrigo Osorio. 2015. “Las encuestas de opinión pública en Chile antes

de 1973.” Latin American Resarch Review 50 (1): 117–39.PNUD. 1998. Informe de Desarrollo Humano en Chile. Las paradojas de la modernización.

Santiago: PNUD.Portales, Felipe. 2000. Chile. Una democracia tutelada. Santiago: Sudamericana.Prothro, James W. and Patricio E. Chaparro. 1974. “Public Opinion and the Movement of

Chilean Government to the Left, 1952–72.” The Journal of Politics 36 (1): 2–43.Riquelme, Alfredo. 1999. “¿Quiénes y por qué `no están ni ahí’? Marginación y/o

automarginación en la democracia transicional. Chile. 1988–1997.” In El modelochileno. Democracia y desarrollo en los noventa, ed. Paul Drake and Ivan Jaksic.Santiago: LOM.

Roberts, Kenneth M., and E. Wibbels. 1999. “Party Systems and Electoral Volatility in LatinAmerica: A Test of Economic, Institutional, and Structural Explanations.” AmericanPolitical Science Review 93 (3): 575–90.

Sartori, Giovanni. 1976. Parties and Party Systems: A Framework for Analysis. New York:Cambridge University Press.

Scully, Timothy R. 1992. Rethinking the Center. Party Politics in Nineteenth- and Twentieth-Century Chile. Stanford: Stanford University Press.

Segovia, Carolina. 2009. “¿Crisis de la política en Chile? Percepciones y valoraciones de lospartidos” In La sociedad de la opinión. Reflexiones sobre encuestas y cambio político endemocracia, ed. Rodrigo Cordero. Santiago: Ediciones Universidad Diego Portales.

Siavelis, Peter. 2002. “Coalitions, Voters and Party System Transformation in Post-Authoritarian Chile.” Government and Opposition 37 (1): 76–105.

Siavelis, Peter. 2004. “Electoral System, Disintegration of Coalitions and Democracy inChile: The End of the Concertacion?” Revista de Ciencia Politica 24 (1): 58–80.


Tironi, Eugenio, and Felipe Agüero. 1999. “¿Sobrevivirá el nuevo paisaje político chileno?”Estudios Públicos 74: 151–68.

Torcal, Mariano and Scott Mainwaring. 2003. “The Political Recrafting of Social Bases ofParty Competition: Chile, 1973–95.” British Journal of Political Science 33 (1): 55–84.

Valenzuela, Arturo. 1978. The Breakdown of Democratic Regimes: Chile. Baltimore: JohnHopkins University Press.

Valenzuela, Arturo. 1990. “Partidos políticos y crisis presidencial en Chile: Proposición de ungobierno parlamentario.” In Hacia una democracia moderna: La opción parlamentaria,ed. Juan Linz, Arend Lijphart, Arturo Valenzuela and Oscar Godoy. Santiago:Universidad Católica de Chile.

Valenzuela, Arturo and J. Samuel Valenzuela, eds. 1976. Chile: Politics and Society. NewBrunswick: Transaction Books.

Valenzuela, J. Samuel. 1995. “The Origins and Transformations of the Chilean PartySystem.” The Helen Kellogg Institute for International Studies at Notre DameUniversity Working Paper #215.

Valenzuela, J. Samuel. 1999. “Respuesta a Eugenio Tironi y Felipe Agüero. Reflexionessobre el presente y futuro del paisaje político chileno a la luz de su pasado.” EstudiosPúblicos 75: 273–90.

Valenzuela, J. Samuel and Timothy R Scully. 1997. “Electoral choices and the party systemin Chile. Continuities and changes at the recovery of democracy.” Comparative Politics29 (4): 511–27.

Valenzuela, J. Samuel, Timothy R. Scully and Nicolás Somma. 2007. “The enduring pres-ence of religion in Chilean ideological positioning and voter options.” ComparativePolitics 40 (1): 1–20.

Winn, Peter, ed. 2004. Victims of the Chilean Miracle. Workers and Neoliberalism in thePinochet era, 1973–2002. Durham: Duke University Press.

Zucco, Cesar. 2010. “Stability Without Roots: Party System Institutionalization in Brazil”Available at SSRN: http://ssrn.com/abstract=2002359 or http://dx.doi.org/10.2139/ssrn.2002359

838 PATRICIO NAVIA

http://ssrn.com/abstract=2002359

http://ssrn.com/abstract=2002359

http://dx.doi.org/10.2139/ssrn.2002359



Documents

It’s the Christian Democrats’ Fault: Declining Political Identification in Chile, 1957–2012